Introduction: Time-limited trials (TLT) are a flexible approach to decision-making about life-sustaining therapies amid uncertainty in critically ill patients. Though increasingly used in ICUs, their conduct varies widely and is shaped by clinician judgment and context. Amidst this wide variation in practice, we have little knowledge about what constitutes an optimal TLT. Understanding their real-world implementation is key to developing effective interventions to support optimal TLTs. Methods: To examine how TLTs are implemented in clinical settings, we conducted a focused ethnography of patients (n = 30) undergoing TLTs in ICUs at five hospitals across the United States. Data sources included direct observations, electronic health record (EHR) notes, audio-recorded family meetings, and interviews with clinicians and family members. Guided by a prior Delphi study that established expert consensus on 17 essential elements of a TLT (“work-as-envisioned”), we developed a process map that served as the basis for a deductive content analysis of each TLT case to assess whether TLT elements were performed fully, partially, or not at all (“work-as-performed”). Results: Across the 30 cases analyzed, none included all 17 work-as-envisioned TLT elements. On average, 72% of elements were addressed per case, with a range of 41.2% to 88.2%. Some elements, such as discussing patient goals, were consistently included (100%), along with planning elements like determining trial duration and post-trial reassessment elements, such as evaluating the patient’s response to therapy (>90% inclusion). Others, such as discussing end-of-trial decisions (57%) and acceptable therapies (63%) appeared less frequently. Physician rotation occurred during 19 TLT cases, and in 2 of these cases, it resulted in post-trial reassessment elements not being completed. Conclusions: Our findings extend the understanding of TLTs by identifying which elements are most and least consistently performed in clinical practice. Further research is needed to better understand the system-level factors driving this variability and to inform strategies for improving fidelity to the TLT process.
Ahmad et al. (Sun,) studied this question.